0/)(2A+-+(&$B1CDC-+A...?+3)@-&")--

GLAST LAT Project Collaboration Meeting Sep 15 – 17, 2003

Benoît Lott Calorimeter Subsystem Status

Calorimeter SubsystemCalorimeter SubsystemStatus and IssuesGamma-ray LargeGamma-ray Large

Area SpaceArea SpaceTelescopeTelescope

Benoît LottCEN Bordeaux-GradignanFranceSlides: Neil Johnson & Eric Grove, NRL


Benoît Lott 2 Calorimeter Subsystem Status

OutlineOutline! Introduction! Subsystem Status! Engineering Model Construction and Test! Flight Manufacturing Status! The Future! Issues and Concerns! Summary! GSI & CERN beam tests



Introduction - Modular DesignIntroduction - Modular Design

4 x 4 Array of Calorimeter Modules

LAT GRID with16 CAL ModulesCAL Module with TEM and PowerSupply mounted to base plate



CsI Detectors + PIN diodes (both ends)

Readout Electronics

Carbon Cell Array Al Cell Closeout

Al EMI Shield Mounting Baseplate

Calorimeter Module OverviewCalorimeter Module Overview

! Electronics boards attached to eachside.

! Electronic readout to connectors atbase of calorimeter.

! Outer wall is EMI shield and providesstructural stiffness as well.

Modular Design4 x 4 array ofcalorimeter modules

Each Module! 8 layers of 12 CsI(Tl) Crystals– Crystal dimensions: 27 x 20 x326 mm– Hodoscopic stacking -alternating orthogonal layers

! Dual PIN photodiode on each end ofcrystals.

! Mechanical packaging – CarbonComposite cell structure



CAL Level III Key RequirementsCAL Level III Key Requirements

Parameter Requirement Verification Expected Performance

Energy Range 20 MeV – 300 GeV 20 MeV – 1 TeV (goal) 5 MeV – 100 GeV, single crystal

Simulation, Beam Tests Required performance ~2 MeV threshold (BOM)

Energy Resolution (1 sigma) < 20% (20 MeV < E < 100 MeV) < 10% (100 MeV < E < 10 GeV) < 6% (10 GeV < E < 300 GeV, incidence

angle > 60 deg)

Simulations and EM and LAT calib unit Beam Tests

Simulations demonstrate required performance

Dead Time < 100 µs per event < 20 µs per event (goal)

Test < 19 µs per event

Low Energy Trigger High Energy Trigger

< 2 µs trigger latency Test < 1 µs

Mass < 1440 kg (90.0 kg/module) Test 1376 kg Power < 65 Watts (conditioned)

(4.05 W/module) Test < 54 Watts (conditioned)

Temperature Range – 10 to +25 C, operational – 20 to +40 C, storage – 30 to +50 C, qualification

Subsystem TV Test 4 cycles, acceptance 12 cycles, qualification

Required performance

Reference: LAT-SS-00018



Status - Loss of CNES Support in FranceStatus - Loss of CNES Support in France! In April, CNES (the French Space Agency) announced theirwithdrawal from planned commitments to the Frenchcontributions to GLAST

! As a consequence, CEA and IN2P3 were forced to reduce theirhardware contributions to CAL.– Purchase of PIN photodiodes and the manufacture of CDEswas transferred from CEA to NRL.– The machining of aluminum and titanium parts of the CALstructure were transferred from IN2P3 to NRL.– Manufacture of composite structure will continue withIN2P3.

! The cost and schedule impact of this change was presented tothe LAT International Finance Committee and the US sponsors.– The revised program and responsibilities have beenapproved and we are in the process of re-baselining theLAT cost and schedule.



Revised CAL Team Roles for FlightRevised CAL Team Roles for FlightOrganization Flight Hardware Responsibility

Sweden

CsI Crystal procurement and acceptance test

France / IN2P3 Mechanical Structure design. Carbon composite cell structures fabrication and test. Elastomer bumpers and cords. MGSE and tooling Finite element and thermal analyses. Beam Test Planning and Support

France / CEA Diode and CDE manufacturing test benches. CDE shipping containers.

Naval Research Lab CAL Subsystem Management, System Engineering, & Mission Assurance. CDE manufacture and test. CAL Electronics Design & Fab, Digital ASIC design, CAL Module Assy & Test, LAT I&T Support

SLAC CAL Analog ASIC Design, EM AFEE PCB layout



CAL StatusCAL Status! CAL Subsystem successfully completed Peer Design Review (Mar ’03)and LAT Critical Design Review (May ’03).

! Engineering Model CAL assembly was completed in March.! Environmental testing of EM CAL was completed in July.– Qualification level vibration tests– Qualification temperature Thermal Vac – 8 cycles.– EMI/EMC testing

! EM CAL was shipped to SLAC for Integration and Test activities inAugust.

! Design revisions for flight production are essentially complete– Modified carbon composite structure manufacturing technique– Modified PIN Diode optical window– New GCFE (analog ASIC) version– Modified CAL base plate to accommodate new CAL-GRID interfacerequirements.

! Delivery of flight components is well underway



Engineering Model CAL ManufacturingEngineering Model CAL Manufacturing

Crystal Detector Element(CDE) AssemblyNRL/CEA

PIN Diode(each end)

CsI Crystal

Optical Wrap

Wireleads

Mechanical StructureFrance (IN2P3/LLR)

Front-End ElectronicsNRL, SLAC

PreElectronics Module (PEM)AssemblyNRL

Module Assemblyand Test, NRL+collab

CsI CrystalsSweden(KTH)

Bond

End Cap

Dual PIN Diodes(DPD)NRL/CEA



CDE ComponentsCDE Components! CDE has four components1. CsI(Tl) crystal2. Two PhotoDiode Assemblies (PDAs)• Hamamatsu S8576-01 Dual PhotoDiode (DPD)• Wire leads, soldered and staked

3. Wrapper• 3M Visual Mirror VM2000 film

4. Two end caps

PIN Diode(each end)

CsI Crystal

Optical Wrap

Wireleads

Bond

End Cap

EM CDEs during wrappingand attachment of end caps



Crystal Production - SwedenCrystal Production - Sweden

Crystal Growing

CrystalCutting

Crystal Polishing

Crystal Mechanical and Optical Acceptance Testing

Amcrys-H, Ukraine

Kalmar Univ, Sweden

a CsI Boule



CsI(Tl) Crystal and PIN PhotodiodeCsI(Tl) Crystal and PIN Photodiode

! CsI(Tl) gives high light yield with PDs andgood stopping power for EM showers– 1536 crystals or ~1200 kg of CsI,each 326 mm x 26.7 mm x 19.9 mm– 100% inspection and test

! PIN Photodiode spectral response wellmatched to CsI(Tl) scintillation– Very small mass, volume, and power– Total 3072 required in LAT CAL– Two diodes to help cover dynamicrange

EM Photodiode Assembly



EM PDA-EM PDA-Crystal BondingCrystal Bonding –– CEA/France CEA/FranceEnd facepolishing

Supporttooling

Primerdeposition

Mold tooling & Glue injection

Mold removalafter 24 hours

Polymerizationtime = 7 days



Support Fixture

EM PDA-EM PDA-Crystal BondingCrystal Bonding –– Swales Swales

EM build:110 CDEs at Swales14 CDEs at Saclay

Hanging xtal in bonding fixture

Priming xtal

Silicone injected,waiting to cure



Composite Structure Composite Structure –– LLR LLR Ecole PolytechniqueEcole Polytechnique

30

60

90

120

150

12:05 13:05 14:05 15:05 16:05 17:05 18:05 19:05Time (h:min)

Temperature (°C)

T consigneT 17T 13T 14T 15

•Each Mandrel Wrappedwith One Pre-Preg Ply

Wrapping of Mandrels Preparation of Layer

• Stacking of Mandrels andLateral Lay-Ups with Inserts• Mechanical Pressure to AddGlobal Plies

Stacking of Layers Closing of Mold

Vacuum Bagging Autoclave Curing Structure Removal Metrology

• Release Film• Breather Felt• Vacuum Bag

•Stacking of Layers, Baseand Top Lay-Ups withInserts

• Temperature 135°C• Pressure 7 bars• Cure Time 4h

• Outer Dimensions• Position of Inserts• Dimension of Cells

• Removal of Layer Frame• Removal of 96 Mandrels• Cleaning

• 4 Side Plates and Cover• Mechanical Stops toControl Outer Dimensions



Pre Electronics Module (PEM) Assembly - NRLPre Electronics Module (PEM) Assembly - NRL

Completed PEM PEM Acceptance Test

Tooling – LLR Ecole Polytechnique

CDE Insertioninto Structure



CAL Electronics CAL Electronics –– NRL/SLAC NRL/SLAC! Read out both ends of CsI crystals inhodoscopic array using PIN photodiodes– 4 printed circuit boards, one on eachvertical face– Large dynamic range (few x 105)– Low noise (~2000 electrons noise)– Low power (~20 mW per crystal end)– Low dead time (20 µs)– Self triggering

! Implementation– Divide dynamic range into two input signals(dual PIN photodiode)– Use 1 custom analog and 1 customdigital ASIC to minimize power– Use COTS 12-bit successive approximationADC on each crystal end to achieve lowdead time.– Sparsify data (zero suppress)

EM AFEE board



AFEE Board InstallationAFEE Board Installation! Route, dress and solder192 diode interconnectwires per AFEE board.

! Functional Test! Stake wires! Install Side Panels

Installation ofEM AFEE boards



EM CAL TestingEM CAL Testing

EM CAL on support standw/ upper lifting fixturesattached

Add picture

EM CAL in TVACMounted on TVAC basecovered w/ thermal shroud

EM CAL in Vibration testMounted on vib table w/accelerometers attached.



EM Calorimeter at SLACEM Calorimeter at SLAC

On arrival at SLAC,EM CAL in andcoming out ofshipping container

Trial insertionof EM CAL intoone-bay grid



EM CDE PerformanceEM CDE Performance! EM CDE build– 110 at Swales Aerospace– 14 at Saclay

! Performance of EM CDEs– Light yield• Big PD within spec– Typical: 8000 e/MeV– EM Spec: >5000 e/MeV

• Small PD within spec– Typical: 1500 e/MeV– EM Spec: >800 e/MeV

Light yield:Big PD

Light yield:Small PD

Saclay and Swales CDEs haveidentical performance



EM Pre-Electronics Module PerformanceEM Pre-Electronics Module Performance! Performance of EM PEM– Assembled PEM with GSECheckout electronics– >5 million muons collected– Data being analyzed with GroundScience Analysis Software system• Muon trajectories imaged• CDE light tapers mapped

Muon energydeposition

Light asymmetry map



EM Module Imaging of EM Module Imaging of Muon Muon TracksTracks! CAL muon position resolution using light tapering along the CsI log.– Light taper maps derived from CAL-only crystal hodoscope– No external hodoscopes! 5 mm resolution (rms)



Flight Manufacturing StatusFlight Manufacturing Status! CsI Crystals– To date Kalmar has received 450 flight CsI xtals from Amcrys H. Ofthese, ~300 have been fully tested and shipped to NRL.

! PDA Manufacturing– NRL has received ~600 flight Dual PIN photodiodes.– PDA manufacturing process tooling and specification have beencompleted. Manufacturing vendor has been selected.– PDAs for pre-qual CDE units are being manufactured at NRL.

! CDEs– More than 30 CDEs have been bonded at Swales for training and toolingtests.– 12 copies of flight CDE bonding tooling have been manufactured.– Remaining Flight CDE manufacturing tooling (38 copies) has beenreleased for manufacture.



Flight Manufacturing Status (2)Flight Manufacturing Status (2)

! Mechanical Structure– Revised, reviewed and released flight machined part drawings –all except base plate which requires review and approval by IPO.– Manufactured Structural Model 1 (SM1) carbon compositestructure using flight-like tooling and autoclave at LLR. 2ndstructure will be made in October.

! AFEE Electronics– Screening, qualification and test boards for AFEE ASICs, ADCs,DACs have been completed.– Flight ASICs have just been received from Mosis. Are now inpackaging.– Revisions to EM AFEE schematic have been completed, newlayout is in progress. Prototype board expected this month.



The FutureThe Future

Nov 14, ‘03 Heavy ion beam test with EM CAL at GSI in Germany

Oct 15, ‘03 Start manufacture of flight CDE Dec 25, ‘03 First PEM Assembled. Feb 20, ‘04 First CAL Module Complete May 28, ‘04 First CAL Module delivered to LAT I&T Nov 02, ‘04 16th CAL Module delivered to LAT I&T Nov 16, ‘04 2 spare modules delivered to SLAC for beam test

calibration



Issues and ConcernsIssues and Concerns! Calorimeter is critically dependent on plastic encapsulatedmicrocircuits (~ 3300 ASICs, ~3000 ADCs and 250 DACs)– Screening and qualification of these will not be completeuntil Jan ’04.

! LAT-wide EMI/EMC design review may require modification /additions to CAL structure– A risk to the delivery schedule



SummarySummary! Engineering Model CAL manufacture and test program hasverified the CAL design and manufacturing processes.– A small number of component / process improvementshave been made for flight.– EM CAL testing continues at SLAC as part of an integratedtower.– The EM will be tested at much higher energies at GSI withheavy ion beams.

! CAL module is expected to meet all its level III requirements.! CAL has begun flight module manufacturing

“Damn the torpedoes, full speed ahead”



}

Use of the ionisation energy loss ofcosmic-ray heavy ionsC, N, O, Mg, Si, Fe

CR energy spectrum

Eloss (E)

“minimum-ionisation” peak

minimum of ionisation

Geomagnetic cutoff

In-orbit calibrationIn-orbit calibration



We need to:

• know the CsI light function L(E,Z), non-linear because ofquenching effects;

• test algorithms for rejecting reaction events (variation of Elossbetween adjacent layers).

CREME96GEANT3

Simulated energy-loss distribution

In-orbit calibration (2)In-orbit calibration (2)



High ionisation density → non-radiative decay channel (“activation-depletion” hypothesis, exciton destruction at activator sites, recombination...)Low energy: Birk’s formula L(E)∝E/(1+kB dE/dx) kB: quenching factorHigh energy:at a given dE/dx, E is higher for greater Z → more d electrons → less quenching

Very scarce data at high energy!NaI(Tl)

Salamon and Ahlen

Quenching effects in Quenching effects in CsICsI



Degrader58Ni Beam

FragmentsTarget

One single beam: 58Ni from 100 to 1700 MeV/nucleon All fragments are produced simultaneously.The spectrometer offers great flexibility!

Experimental setup at GSIExperimental setup at GSI



Why do a beam test at CERN?Why do a beam test at CERN?

The LAT will cover the 20 MeV-300 GeV range, but the calibration at SLAC will be limited to 30 GeV. Wouldn’t be data at higher energies be useful? Although rare, photons with E>10 GeV will (arguably) be the most interesting, the 10 GeV-100 GeV domain has barely been explored.

Only little ressource is needed to get these data.

Our calorimeter is fairly thin: on-axis photons deposit only ~40% of their energywithin the CAL: energy reconstruction for those is based on profile fitting.Surprisingly few data on shower profiles in the literature.

CERN is a good place:GLAST is an “accepted experiment” at CERN.Different beams (secondary+tertiary) are available with a great flexibility: electrons, pions,muons with energies up to 150 (300) GeV, down to 10 GeV CERN shut down in 2005. Beam available for GLAST in 2004 (?)



Experimental MethodExperimental Method

Duration : 1 week of beam time on the H6 beam line (Aug 7-13)Detectors: 48 CDEs (AMCRYS): 8 Bordeaux, 15 CEA, 25 Kalmar+CEA arranged in 8 layers of 6 15 CDEs (BTEM) from NRL arranged in 3 layers of 5 positioned on a moving table + rotating table (48)Electronics: 222 channels preamp + 2 sets of shaping amplifiers low-gain “x1”, high-gain “x20” commercial VME ADCs (CAEN 785)Localization: 2 X-Y Silicon chambers (Trieste)Trigger: 2x2x0.2 cm3 plastic scintillator located 6 m in front of the detectorBeams: electrons at 10-20-50-80-120-150 GeV, 20 GeV muons, 20 GeV pionsCount rate: 100 Hz“Converter”: 15x15 cm2 Pb sheets of various thicknesses (0-16 X0, 1.3 X0)People: Bordeaux: 7, Kalmar: 3, CEA: 2, Trieste:2



SetupSetup



All we have for absolute calibration are muons E-deposits.Can we use the muon data alone (Edep~12 MeV) to establish the calibration upto tens of GeV?

Procedure:Muons in Big Diodes (BD) corrected for attenuation*Conversion Slope for BDs with high-gain amplifiersPulser: relative gains between low- and high-gain amplifiersConversion slopes for Big Diodes with low-gain amplifiersSmall-diode vs Big-diode correlation using beam data:Conversion slopes for Small Diodes

*attenuation coefficients: Left/Right dependence on position

Energy CalibrationEnergy Calibration



MC GEANT3data

high-gain amplifiers

Energy calibration with a 20 Energy calibration with a 20 GeV muon GeV muon beambeam

2002 experiment



Longitudinal shower profilesLongitudinal shower profiles

50 GeV 1.3 X0

80 GeV 1.3 X0

120 GeV 0 X0

120 GeV 4 X0

120 GeV 8 X0



Last layerLast layer correlationcorrelation

100 evts….

data GEANT4



Light collection asymmetryLight collection asymmetry



Position measurementPosition measurement

Documents

0/)(2A+-+(&$B1CDC-+A...?+3)@-&")--